Explore the vast range of titles available.

Using polyaspartic acid (PAsp) and bentonite (BT) as the main raw materials, a new type of degradable soil water retaining agent (PAsp-AA/BT) was synthesized by microwave radiation. The optimum synthesis conditions and comprehensive properties of PAsp-AA/BT were discussed and the structure and surface characteristics of PAspsp-AA/BT were characterized by FTIR, SEM, XRD and TGA in the paper. The results showed that the optimum synthesis conditions of PAsp-AA/BT were as follows: the dosages of polyaspartic acid (PAsp), bentonite (BT), initiator potassium persulfate, crosslinking agent N , N ′-methylene bisacrylamide was 5, 3, 0.3, 0.03%, respectively, the neutralization degree of acrylic acid was 75%, and the microwave power was 490W. Under this condition, the absorption ratio of the synthesized PAspsp-AA/BT in deionized water and 0.9% NaCl solution was 953 and 164 g/g, respectively. The synthesized PAsp-AA/BT had a high water absorption rate, good water retention and repeated water absorption, and the degradation rate in soil within 30 days reached 32.75%, with good degradation effect. The analysis of SEM, FT-IR, XRD and TGA showed that: the surface of PAsp-AA/BT was rough and had obvious pore structure, which was conducive to the diffusion of water molecules; polyaspartic acid, bentonite and acrylic acid were polymerized; the cross-linking structure was formed between polyaspartic acid, bentonite and acrylic acid; the product of PASP-AA/BT had good thermal stability. This study provides a new soil water retaining agent, which is helpful for the better development of soil water retaining agent research.

This paper demonstrates the one-step hydrothermal synthesis of a tin dioxide (SnO 2 )-decorated reduced graphene oxide (RGO) hybrid nanocomposite, which was drop-casted on a polyimide substrate as a humidity sensor. The as-synthesized hybrid was characterized in terms of its nanostructural, morphological and compositional features by SEM, XRD and nitrogen sorption. The humidity sensing properties of the presented RGO/SnO 2 hybrid nanocomposite, such as repeatability, stability, response-recovery characteristics, were investigated by exposing it to a broad humidity range of 11–97% RH at room temperature. As a result, the sensor demonstrated a high sensitivity, a good repeatability, an acceptable linearity, a fast response/recovery characteristic and high long-term stability over a full humidity range measurement, indicating the unique advantages of one-step hydrothermal synthesis for sensor fabrication. The possible and proposed sensing mechanism for the sensor is mainly attributed to a humidity-induced transfer of charge carriers occuring at the interfaces and the swelling effect of RGO.

Key message In this manuscript, authors reviewed and explore the information on beneficial role of phytohormones to mitigate adverse effects of heavy metals toxicity in plants. Global farming systems are seriously threatened by heavy metals (HMs) toxicity, which can result in decreased crop yields, impaired food safety, and negative environmental effects. A rise in curiosity has been shown recently in creating sustainable methods to reduce HMs toxicity in plants and improve agricultural productivity. To accomplish this, phytohormones, which play a crucial role in controlling plant development and adaptations to stress, have emerged as intriguing possibilities. With a particular focus on environmentally friendly farming methods, the current review provides an overview of phytohormone-mediated strategies for reducing HMs toxicity in plants. Several physiological and biochemical activities, including metal uptake, translocation, detoxification, and stress tolerance, are mediated by phytohormones, such as melatonin, auxin, gibberellin, cytokinin, ethylene, abscisic acid, salicylic acid, and jasmonates. The current review offers thorough explanations of the ways in which phytohormones respond to HMs to help plants detoxify and strengthen their resilience to metal stress. It is crucial to explore the potential uses of phytohormones as long-term solutions for reducing the harmful effects of HMs in plants. These include accelerating phytoextraction, decreasing metal redistribution to edible plant portions, increasing plant tolerance to HMs by hormonal manipulation, and boosting metal sequestration in roots. These methods seek to increase plant resistance to HMs stress while supporting environmentally friendly agricultural output. In conclusion, phytohormones present potential ways to reduce the toxicity of HMs in plants, thus promoting sustainable agriculture.

A high-performance methane gas sensor based on nickel oxide (NiO)/reduced graphene oxide (rGO) nanocomposite film was reported in this paper. The hydrothermal synthesized NiO/rGO hybrid nanocomposite was fabricated on a ceramic tube as sensing film. The nanostructures of the NiO/rGO nanocomposite film were characterized by scanning electron microscopy, X-ray diffraction and transmission electron microscope. The methane gas sensing behaviors of the sensor samples were investigated by exposing to various concentration of methane gas at different operating temperature. As a result, the presented sensor exhibited high-response, good repeatability and acceptable selectivity toward methane gas detection. The possible gas sensing mechanism of the proposed sensor was attributed to the Fermi energy band between rGO sheets and NiO nanoparticles. This observed results highlight the hydrothermal synthesized NiO/rGO nanocomposite film can be used as a candidate material for constructing methane sensors, given its simple process, practical usability and cost effectiveness.

Recent advances in machine learning (such as ChatGPT and DALL·E) have significantly changed people’s lives. At the same time, as people become increasingly reliant on immensely these powerful models, there is a growing recognition that the ethical issues associated with these models also need to be addressed, among which algorithmic fairness and data privacy are two of the most concerning aspects. While algorithmic fairness ensures that a resulting ML model does not favor a particular data group over others (e.g., male job applicants over female, rich patients over poor, etc.), data privacy ensures that the sensitive information about the underlying training dataset is not revealed during model training and after model deployment.In this thesis, we study the fairness and privacy issues of machine learning over distributed data, as we are in a world where data are often born in different places and possessed by different parties (e.g., edge devices, local governments, and mobile service providers that partition the overall user population). Addressing privacy and fairness issues is more challenging in distributed settings, compared to the traditional central setting. While a central server could formalize fairness and privacy into the learning objective and solve the problem by running standard optimization algorithms on the overall data, the straightforward application to the distributed setting could lead to privacy issues. Instead, we focus on the most popular learning paradigm for distributed data, namely, Federated Learning (FL), where different parties share their model parameters instead of their local during the training process.We approach the privacy and fairness issues of federated learning for distributed data through a series of systematic studies. We first provide an efficient algorithm for parties to audit the privacy risks in FL as a precautionary measure for data privacy. Next, we reveal the inherent unfairness issue in FL– parties that are biased toward certain groups may encode this bias in model parameters and propagate such bias through model aggregation, an elementary operation taken by parties in FL. We also study the inherent trade-offs between privacy and fairness and demonstrate that fairness may come with the cost of data privacy. Finally, recent large language models (LLMs) are trained on vast datasets, which may include sensitive information contributed by various parties. However, existing privacy auditing frameworks designed for traditional machine learning models do not accurately estimate the risks associated with LLMs. To address this issue, we have developed new auditing algorithms specifically tailored to the unique characteristics of LLMs. Building on our auditing framework, we plan to further investigate the privacy risks involved in training LLMs using federated learning methods in the future.

This paper demonstrates a hydrogen gas sensor based on palladium–zinc oxide–reduced graphene oxide (Pd–ZnO/RGO) ternary hybrid via hydrothermal route. The nanostructures of Pd–ZnO/RGO hybrid nanocomposites were fully characterized by scanning electron microscopy and X-ray diffraction, confirming its successful preparation. The Pd–ZnO/RGO hybrid film sensor was investigated by exposing to hydrogen gas with concentration in a wide range of from 1 ppb to 500 ppm. The experiment results exhibit that the presented sensor has a hypersensitive response, quite short response/recovery times, and achieves ppb-level ultralow detection toward hydrogen at low working temperature. The underlying sensing mechanism of the Pd–ZnO/RGO hybrid toward hydrogen was attributed to the synergistic effect of the ternary hybrid nanomaterials and the modulation of the depletion layer width at the interfaces. This work indicates that the ternary Pd–ZnO/RGO hybrid is a candidate for constructing high-performance hydrogen gas sensors for various applications.

We show that participating in federated learning can be detrimental to group fairness. In fact, the bias of a few parties against under-represented groups (identified by sensitive attributes such as gender or race) can propagate through the network to all the parties in the network. We analyze and explain bias propagation in federated learning on naturally partitioned real-world datasets. Our analysis reveals that biased parties unintentionally yet stealthily encode their bias in a small number of model parameters, and throughout the training, they steadily increase the dependence of the global model on sensitive attributes. What is important to highlight is that the experienced bias in federated learning is higher than what parties would otherwise encounter in centralized training with a model trained on the union of all their data. This indicates that the bias is due to the algorithm. Our work calls for auditing group fairness in federated learning and designing learning algorithms that are robust to bias propagation.

Statistical measures for group fairness in machine learning reflect the gap in performance of algorithms across different groups. These measures, however, exhibit a high variance between different training instances, which makes them unreliable for empirical evaluation of fairness. What causes this high variance? We investigate the impact on group fairness of different sources of randomness in training neural networks. We show that the variance in group fairness measures is rooted in the high volatility of the learning process on under-represented groups. Further, we recognize the dominant source of randomness as the stochasticity of data order during training. Based on these findings, we show how one can control group-level accuracy (i.e., model fairness), with high efficiency and negligible impact on the model's overall performance, by simply changing the data order for a single epoch.

The present study was undertaken to investigate the effects of Agaricus blazei murill polysaccharides on morphology of hippocampus tissue and expression of ASIC1a and ASIC2b mRNA in lead-poisoning rats.Forty SD rats were divided into five groups：control group,lead group,experimental group 1,2,3.After 60d of treatments,hippocampus tissues were observed by HE staining and the expression levels of ASIC1a and ASIC2b mRNA were determined by the flouoreescent quantitative real-time polymerase chain reaction（FQRT-PCR）.The results showed that after lead induction,the number of hippocampus neuron increased,cell gap increased,the chromatin was stained lightly;after administration of ABMP,the number of the hippocampus neuron increased as the dose of ABMP rose,cell gap decreased,the chromatin was stained darkly.The expression level of ASIC1a mRNA in hippocampus of rats in ABMP group decreased extremely significantly compared with that in the model group（P〈0.01）.Lead exposure and ABMP exhibited no significant effects on the expression level of ASIC2b mRNA.ABMP possesses a protective effect on nervous system of rats exposed to lead by combining with lead.

Watermarking is a technique used to embed a hidden signal in the probability distribution of text generated by large language models (LLMs), enabling attribution of the text to the originating model. We introduce smoothing attacks and show that existing watermarking methods are not robust against minor modifications of text. An adversary can use weaker language models to smooth out the distribution perturbations caused by watermarks without significantly compromising the quality of the generated text. The modified text resulting from the smoothing attack remains close to the distribution of text that the original model (without watermark) would have produced. Our attack reveals a fundamental limitation of a wide range of watermarking techniques.